Policymakers today face impossible choices. They are working from models that send mixed signals and are fed by poor-quality data. Italy, which has the most advanced epidemic in Europe, does not know how many infections are undetected or deaths are unreported. Even the number of tests does not seem to align with the number of people who have been tested.
Comparing data between countries is even more perilous, given radically different standards for its collection. And with radically divergent estimates on COVID-19’s CFR, R0 (and the linked threshold for herd immunity), likely duration of immunity, and risks of second and third waves, decisions about when and how to loosen lockdowns will prove extremely hard to make.
The pathways to reducing uncertainty about the public health fundamentals – robust estimates of rates of excess deaths and analysis of why they vary in different settings, mass serological testing and monitoring of antibody response – don’t exist and won’t for a while.
Meanwhile, projections of economic growth are unreliable in normal times and little better than astrology during this unprecedented crisis. Just a month ago, many analysts were confidently predicting a V-shaped recovery. Now the IMF expects a downturn as bad as the Great Depression, but is suitably cautious when saying the costs “could be around 9 trillion dollars”.
What is clear is that many secondary crises are only just brewing. Global supply chains are in trouble. Crops are rotting in the fields. Hunger is increasing in rich and poor countries. The IMF graph we have most confidence in? The one that says that no-one really knows anything at the moment:
Source: Ahir, Bloom, and Furceri (2018) | worlduncertaintyindex.com
We’ve written (if tweeting counts as writing) about the need for adaptive decision making. And about how our political systems will fail in this crisis if they don’t recognise what their own comparative advantage is (political trade-offs) and how to defer to other sorts of expertise when needed.
We’re not alone in pursuing this line. The RSA has done great work on the science:policy interface. David Nabarro has written compellingly on people-centred systems leadership and on the need to learn fast. Dr. Rama Dieng’s thread of articles from Global South feminists has highlighted the need for a diversity of perspectives – for example, by clarifying how social distancing simply won’t work in certain communities, such as the slums of Lagos.
The only thing that is certain is that the uncertainty will continue. So how do you make good decisions when you’re playing whack-a-mole? Here are four recommendations for governments to improve their decision making.
1. Form an independent red-team
This team should be tasked with reviewing government policy ideas and providing a formal, rapid, and public critique prior to a final decision. Red teams give you an ability to check and re-check assumptions and help avoid the natural groupthink that occurs under pressure.
The average age should be younger than the top tier of government decision makers and much more diverse. Include an ethicist, a grassroots organiser or frontline health worker, a person with disabilities, and people from groups with high COVID-19 risk, such as ethnic minorities, the elderly, or those with pre-existing conditions.
No more than 12 members. A rotating chair. Two researchers and a professional facilitator. Access to expertise as needed. Formal submissions to governments that are limited to four pages and no more.
2. Empower a ‘mole-spotting’ unit
There’s too much happening and too much information that is not consolidated into a form that helps decision makers.
So, each week make sure your main decision-making body (the Cabinet or equivalent) receives a briefing on an aspect of the crisis that is not yet on their radar. This should cover an emerging challenge, potential solutions, and make recommendations for how to find out more.
The mole spotters should also think structurally about knowledge needs. What research needs to be commissioned now – or which findings systematically reviewed – to make sure that next month’s questions can be better answered. This both helps decision makers have peripheral vision and constantly improve decision quality.
3. Embrace foresight to manage risks
In a crisis, we’re prey to false dichotomies (“save the airlines” or “no airlines”) and to over-estimating the pace of change but underestimating the magnitude. We make decisions in a hurry that have massive path dependencies – shutting off possible futures and large parts of society from any good future at all.
Counter this by hacking out space for some longer-term bets. Identify a handful of initiatives to start today that you expect to have outsized impact on your country’s position in one, five, and fifty years’ time. This is critical to offering hope to societies that we can, as the poet has it, make this place beautiful.
How to build a climate-viable future for air travel as the industry arises from the ashes, for example.
Or how to give your society’s most alienated groups a sense of agency in the emergency response in a way that will start the process of meaningfully addressing their grievances and promoting their inclusion.
But then also revisit foresight scenarios, regularly, to understand how to keep adapting to keep those bets on track. This helps decision makers continually adapt to emerging risks and shifting probabilities – a core condition of uncertainty.
4. Build in real feedback loops
Nobody likes to fail. Being seen to fail is worse. And then we create powerful disincentives for reporting failure. Most senior public servants are spending more time scribbling down private memoranda to defend themselves at a future commission of enquiry than in learning from what’s going wrong in the here and now.
Survey your frontline workers every week – healthcare, police, retail employees, etc. Any reputable polling organisation could have a panel up and running within a few days and the savings from making fewer mistakes will be orders of magnitude greater than the costs.
And pay some experts to create a safe and independent space to review government decision making, help wind down projects safely, and provide steady feedback on what has and hasn’t worked.
This is how experimentation feeds learning and defers to frontline expertise.
Rahul Chandran writes here in his personal capacity.
